Paper: arXiv 2503.13193

Abstract

We introduce the deep multi-FBSDE method for robust approximation of coupled forward-backward stochastic differential equations (FBSDEs), focusing on cases where the deep BSDE method of Han, Jentzen, and E (2018) fails to converge. To overcome the convergence issues, we consider a family of FBSDEs that are equivalent to the original problem in the sense that they satisfy the same associated partial differential equation (PDE). Our algorithm proceeds in two phases: first, we approximate the initial condition for the FBSDE family, and second, we approximate the original FBSDE using the initial condition approximated in the first phase. Numerical experiments show that our method converges even when the standard deep BSDE method does not.

Complexity vs Empirical Score

  • Math Complexity: 9.0/10
  • Empirical Rigor: 4.0/10
  • Quadrant: Lab Rats — theoretically deep, empirically untested

Why this score: The paper is highly theoretical, employing advanced stochastic calculus, FBSDE theory, and rigorous numerical analysis, but lacks backtest-ready code or heavy implementation focus, showing only controlled numerical experiments.

Research Flowchart

  flowchart TD
  A["Research Goal<br>Robust Deep Learning for Coupled FBSDEs"] --> B["Data/Input<br>FBSDE-PDE System Parameters"]
  B --> C["Methodology: Two-Phase Training"]
  C --> D["Phase 1: Approximate<br>Initial Condition via Equivalent FBSDE Family"]
  D --> E["Phase 2: Approximate<br>Original FBSDE using Initial Condition"]
  E --> F["Computational Process<br>Monte Carlo Simulation & Neural Network Optimization"]
  F --> G["Key Finding<br>Convergence Achieved where Standard Deep BSDE Method Fails"]

Cited References (144)

  1. Three algorithms for solving high-dimensional fully coupled FBSDEs through deep learning (2020)
  2. Convergence of the Markovian iteration for coupled FBSDEs via a differentiation approach — → 2504.02814
  3. Neural Network for CVA : Learning Future Values (2017) — link
  4. Integrated structural approach to Credit Value Adjustment (2019) — link
  5. Partial Differential Equation Representations of Derivatives with Bilateral Counterparty Risk and Funding Costs (2011)
  6. In the balance (2011)
  7. Cross Currency Valuation and Hedging in the Multiple Curve Framework (2021) — link
  8. A NOTE ON THE SELF-FINANCING CONDITION FOR FUNDING, COLLATERAL AND DISCOUNTING (2015) — link
and 136 more
  1. A note on the self-financing condition for funding, collateral and discounting (2015)
  2. Arbitrage-free bilateral counterparty risk valuation under collteralization and application to credit default swaps (2014) — link
  3. Arbitrage-free valuation of bilateral counterparty risk for interest-rate products: Impact of volatilities and correlations (2011)
  4. Risk Neutral Pricing of Counterparty Risk
  5. Counterparty Risk in Derivatives and Collateral Policies: The Replicating Portfolio Approach
  6. Swap Rates and Credit Quality (1996) — link
  7. Nonlinear valuation with XVAs: two converging approaches (2022)
  8. Funding, collateral and hedging: uncovering the mechanics and the subtleties of funding valuation adjustments (2012) — link
  9. Funding valuation adjustment: a consistent framework including cva, dva, collateral, netting rules and re-hypothecation (2011) — link
  10. When Capital Is a Funding Source: The XVA Anticipated BSDEs (2020)
  11. Arbitrage-Free XVA (2018)
  12. Analysis of non-linear approximated value equation under multiple risk factors and stochastic intensities (2023)
  13. Backward stochastic differential equations (2017)
  14. A review of tree-based approaches to solve forward-backward stochastic differential equations (2021)
  15. Deep multi-step mixed algorithm for high dimensional non-linear PDE s and associated BSDE s — → 2308.14487
  16. Nonlinear consistent valuation of CCP cleared or CSA bilateral trades with initial margins under credit, funding and wrong-way risks (2014)
  17. Dangers of Bilateral Counterparty Risk: the fundamental impact of closeout conventions (2010) — link
  18. Modeling Default Risk (2002)
  19. Numerical method for backward stochastic differential equations (2002)
  20. Statistical Learning of Value-at-Risk and Expected Shortfall (2024)
  21. Regression quantiles (1978)
  22. Learning value-at-risk and expected shortfall (2022) — link
  23. Approximation error analysis of some deep backward schemes for nonlinear PDE s (2022)
  24. Adaptive Multilevel Stochastic Approximation of the Value-at-Risk — → 2408.06531
  25. A F ourier cosine method for an efficient computation of solutions to BSDEs (2015)
  26. Efficient numerical F ourier methods for coupled forward–backward SDE s (2016)
  27. Numerical F ourier method and second-order Taylor scheme for backward SDE s in finance (2016)
  28. Solving forward-backward stochastic differential equations explicitly - A four step scheme (1994)
  29. Error expansion for the discretization of backward stochastic differential equations (2007)
  30. Solving BSDE with adaptive control variate (2010)
  31. On multilevel P icard numerical approximations for high-dimensional nonlinear parabolic partial differential equations and high-dimensional nonlinear backward stochastic differential equations (2019)
  32. A forward scheme for backward SDE s (2007)
  33. Discrete-time approximation and M onte- C arlo simulation of backward stochastic differential equations (2004)
  34. Stochastic grid bundling method for backward stochastic differential equations (2019)
  35. A probabilistic numerical method for fully nonlinear parabolic PDE s (2011)
  36. Least-squares M onte C arlo for backward SDE s (2012)
  37. On the M onte C arlo simulation of BSDE s: A n improvement on the M alliavin weights (2010)
  38. Stratified regression M onte- C arlo scheme for semilinear PDE s and BSDE s with large scale parallelization on GPU s (2016)
  39. Counterparty risk and funding: immersion and beyond (2016) — link
  40. XVA analysis from the balance sheet (2021)
  41. An Explicit Scheme for Pathwise XVA Computations (2024) — link
  42. Pathwise CVA regressions with oversimulated defaults (2023)
  43. Numerical methods for backward stochastic differential equations: A survey (2021) — link
  44. Convergence of the deep BSDE method for stochastic control problems formulated through the stochastic maximum principle — → 2401.17472
  45. Deep backward schemes for high-dimensional nonlinear PDE s (2020)
  46. Solving high-dimensional partial differential equations using deep learning (2018)
  47. Deep x VA solver– A neural network based counterparty credit risk management framework (2023) — link
  48. A posteriori error estimates for fully coupled McKean–Vlasov forward-backward SDEs (2024)
  49. Generalized convergence of the deep BSDE method: a step towards fully-coupled FBSDE s and applications in stochastic control (2024) — link
  50. A deep solver for BSDE s with jumps (2022) — link
  51. Multi-Layer D eep x VA : Structural C redit M odels, M easure C hanges and C onvergence A nalysis — → 2502.14766
  52. A multistep scheme to solve backward stochastic differential equations for option pricing on GPUs (2021)
  53. Backward stochastic differential equations and quasilinear parabolic partial differential equations (2005)
  54. Probabilistic interpretation for systems of quasilinear parabolic partial differential equations (1991)
  55. Bilateral counterparty risk under funding constraints—Part I: Pricing (2015)
  56. Bilateral counterparty risk under funding constraints—Part II: CVA (2015)
  57. Collateralization, Nonlinear Valuation Under and Risk, Credit (2016)
  58. Positive XVAs (2022)
  59. A unified approach to xVA with CSA discounting and initial margin (2021)
  60. XVA Principle, Nested Monte Carlo Strategies, and GPU Optimizations (2018) — link
  61. Mild to classical solutions for XVA equations under stochastic volatility (2024)
  62. Nonlinear valuation under credit, funding, and margins: Existence, uniqueness, invariance, and disentanglement (2019)
  63. A backward differential deep learning-based algorithm for solving high-dimensional nonlinear backward stochastic differential equations — → 2404.08456
  64. A numerical Fourier cosine expansion method with higher order Taylor schemes for fully coupled FBSDEs (2025) — link
  65. Convergence of the deep BSDE method for coupled FBSDE s (2020)
  66. Convergence of the Deep BSDE method for FBSDE s with non- L ipschitz coefficients (2021)
  67. A proof that rectified deep neural networks overcome the curse of dimensionality in the numerical approximation of semilinear heat equations (2020)
  68. Analysis of the generalization error: E mpirical risk minimization over deep artificial neural networks overcomes the curse of dimensionality in the numerical approximation of B lack– S choles partial differential equations (2020)
  69. D NN expression rate analysis of high-dimensional PDE s: Application to option pricing (2021)
  70. A proof that artificial neural networks overcome the curse of dimensionality in the numerical approximation of B lack– S choles partial differential equations (2023) — link
  71. A proof that deep artificial neural networks overcome the curse of dimensionality in the numerical approximation of K olmogorov partial differential equations with constant diffusion and nonlinear drift coefficients (2021)
  72. Solving the K olmogorov PDE by means of deep learning (2021)
  73. Machine learning approximation algorithms for high-dimensional fully nonlinear partial differential equations and second-order backward stochastic differential equations (2019)
  74. Deep splitting method for parabolic PDE s (2021)
  75. Asymptotic expansion as prior knowledge in deep learning method for high dimensional BSDE s (2019)
  76. Deep primal-dual algorithm for BSDE s: Applications of machine learning to CVA and IM (2017)
  77. Forward–backward stochastic neural networks: deep learning of high-dimensional partial differential equations (2024)
  78. Learning L ocomotion C ontrollers for W alking U sing D eep FBSDE (2021) — link
  79. Machine learning for semi linear PDE s (2019)
  80. The O ne S tep M alliavin scheme: new discretization of BSDE s implemented with deep learning regressions (2024)
  81. A novel pricing method for European options based on Fourier-cosine series expansions (2009)
  82. Convergence of a robust deep FBSDE method for stochastic control (2023)
  83. An overview on deep learning-based approximation methods for partial differential equations (2020) — link
  84. Counterparty credit risk in a multivariate structural model with jumps (2015)
  85. Deep learning for CVA computations of large portfolios of financial derivatives (2021)
  86. A deep learning approach for computations of exposure profiles for high-dimensional Bermudan options (2021)
  87. Learning deep stochastic optimal control policies using forward-backward SDE s (2019) — link
  88. Deep FBSDE C ontroller for A ttitude C ontrol of H ypersonic A ircraft (2021)
  89. Strong approximations of BSDEs in a domain (2009)
  90. Time discretization and Markovian iteration for coupled FBSDEs (2008)
  91. Pricing barrier options with deepBSDEs (2020) — link
  92. Barrier options and Greeks: Modeling with neural networks (2023)
  93. A deep neural network algorithm for semilinear elliptic PDEs with applications in insurance mathematics (2020)
  94. Convergence of the backward deep BSDE method with applications to optimal stopping problems (2023)
  95. A control method for solving high-dimensional Hamiltonian systems through deep neural networks (2021) — link
  96. Deep BSDE-ML L earning and I ts A pplication to M odel- F ree O ptimal C ontrol (2022) — link
  97. Deep learning approximation for stochastic control problems (2016) — link
  98. Backward forward stochastic differential equations (1993)
  99. Characterization of bistability for stochastic multistep methods (2012)
  100. Two-sided error estimates for the stochastic theta method (2010)
  101. Convergence of the Euler scheme for a class of stochastic differential equation (2002)
  102. Stochastic differential equations (1992)
  103. Rate of convergence of the Euler approximation for diffusion processes (1991)
  104. Weak rate of convergence of the E uler– M aruyama scheme for stochastic differential equations with non-regular drift (2017)
  105. Introduction to stochastic control theory (2012)
  106. Approximations by superpositions of a sigmoidal function (1989)
  107. Activation functions: Comparison of trends in practice and research for deep learning (2018) — link
  108. On the Pricing of Corporate Debt: The Risk Structure of Interest Rates (1974)
  109. Temporal difference learning for high-dimensional PIDEs with jumps (2024)
  110. Deep learning algorithms for FBSDEs with jumps: Applications to option pricing and a MFG model for smart grids (2024) — link
  111. Valuing A merican options by simulation: a simple least-squares approach (2001)
  112. On the pricing of corporate debt: The risk structure of interest rates (1974)
  113. Valuing corporate securities: Some effects of bond indenture provisions (1976)
  114. Valuation adjustments: A balanced perspective (2011)
  115. Counterparty Risk and Funding: A Tale of Two Puzzles (2015)
  116. Counterparty risk and funding: A tale of two puzzles (2014)
  117. XVA: Theory, Implementation, and Practice (2019)
  118. A change-of-measure approach for wrong-way risk (2018)
  119. Disentangling wrong-way risk: pricing CVA via change of measures and drift adjustment (2018) — link
  120. Augmenting Structural Credit Models with Stochastic Recovery (2022)
  121. Deep learning-based numerical methods for high-dimensional partial differential equations (2017)
  122. Deep BSDE methods for high-dimensional nonlinear PDEs: A review (2019)
  123. Solving Coupled Nonlinear Forward-backward Stochastic Differential Equations: An Optimization Perspective with Backward Measurability Loss (2023) — link
  124. Basel III: A Global Regulatory Framework for More Resilient Banks and Banking Systems (2010)
  125. Probabilistic Theory of Mean Field Games with Applications I (2018)
  126. Deep optimal stopping (2019)
  127. Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations (2021)
  128. Convergence of a Deep BSDE solver with jumps — → 2501.09727
  129. Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations (2019)
  130. Deep backward schemes for high-dimensional nonlinear PDEs (2020)
  131. $\Gamma$-convergence and optimal control problems (1982)
  132. Variational analysis (2009)
  133. On the time discretization of stochastic optimal control problems: the dynamic programming approach (2019)
  134. Adam: A method for stochastic optimization (2014) — link
  135. Backward deep BSDE methods and applications to nonlinear problems (2020) — link
  136. Deep Learning algorithms for solving high dimensional nonlinear Backward Stochastic Differential Equations (2020) — link

7 of 144 cited works are in the corpus; resolved via arXiv source bundles, Semantic Scholar and OpenAlex. Corpus-wide: 2518 linked papers.